1. Field of the Invention
The present invention relates to a projection method and a projection system and a mask therefor, and more particularly, to a projection method for exposing an object and a projection system and a mask therefor using a modified illumination method that is used for a lithography process of a semiconductor device.
2. Prior Art
A resolution of 0.3 .mu.m and an adequate depth of focus (DOF) are necessary in order to manufacture ULSI chips having an integration of 64M bits or more, and accordingly, many new techniques are under development so as to form patterns of less than one half micron. Examples of these include a method using an excimer laser wherein a short wavelength of light is used, an exposure method using a phase shift mask, and a modified illumination method such as a tilted illumination method.
It is known that the resolution (R) and the DOF of optical projection exposure are proportional to the exposure wavelength (.lambda.) and inversely proportional to the numerical aperture (NA) of a lens, as shown in the following Rayleigh's equation: ##EQU1##
Here, k.sub.1 and k.sub.2 are process factors, i.e. coherence factors (expressed as .sigma. which is a value obtained by dividing NA.sub.c of an illumination system lens by NA.sub.0 of a projection optical system) of an illumination system, and are generally known as being from 0.5 to 0.8. When a coherence factor is not zero, the values of k.sub.1 and k.sub.2 cannot be known precisely since resolution is largely affected by process ability.
When a pattern having a size below 0.3 .mu.m is to be formed using the current projection exposure method, a DOF larger than 1.6 .mu.m is necessary, due to the large step height of the processed wafer as well as other errors.
In the current stepper, when an exposure is performed by using an i-line, the resolution (R) and DOF are 0.47 .mu.m and 1.46 .mu.m, respectively, since .lambda. is 0.365 .mu.m, and k.sub.1 and k.sub.2 are 0.65 and 1.0, respectively.
In order to obtain a higher resolution, NA should be larger or .lambda. should be shorter. When using the conventional KrF excimer laser stepper having an NA value of 0.54 (where .lambda.=0.248 .mu.m), the resolution (R) is 0.3 .mu.m while the DOF is 0.85 .mu.m. Accordingly, since the resolution is improved and the DOF is decreased, patterns below 0.3 .mu.m are very difficult to form by applying the current projection exposure method, even though a KrF excimer laser is used.
Additionally, the method wherein a phase shift mask is used is expensive due to the manufacturing costs of the phase shift mask. Also, the mask's manufacture is difficult.
Recently, a method (hereinafter called a "modified illumination method") has been suggested for exposing a resist by a tilted illumination, wherein an illumination system having a filter attached between the fly's eye lens and the condenser lens is used (see "New imaging technique for 64M DRAMs" by N. Shiraishi, S. Hirukuwa, Y. Takeuchi, and N. Magome, Proceedings of SPIE, Vol.1674, Optical/laser Microlithography, p741, 1992).
The conventional modified illumination method is explained by referring to FIGS. 1 and 2.
FIG. 1 illustrates a structure of a modified illumination system of the conventional projection exposure system, and FIG. 2 illustrates the shape of a filter attached to the illumination system of FIG. 1.
The conventional modified illumination system is comprised of a conventional illumination apparatus which has a light source 1, a projection lens (not shown), a fly's eye lens 2 and a condenser lens 4, with a filter 3 attached thereto. Filter shapes are shown in FIGS. 2A and 2B. Here, FIG. 2A shows an annular illumination system, while FIG. 2B shows a quadruple illumination system. As shown in FIG. 1, vertical incident components of incident light are blocked by the filter as described above, and only oblique incident components of light illuminates mask 5. This is called a tilted illumination method.
Referring to FIGS. 3 and 4, the tilted illumination will be explained in more detail. FIG. 3 is a schematic diagram showing a conventional projection exposure method and FIG. 4 is a schematic diagram showing an projection exposure method using the above tilted illumination method.
Generally, the illumination light from the light source is limited by filter 3 at the exit surface of the fly's eye lens, which coincides with the Fourier transform plane of the mask through the condenser lens. According to the conventional projection exposure method shown in FIG. 3(a), the distribution of illumination light on the Fourier transform plane falls within a circular area. Zero-order diffracted light travels along the optical axis (vertically incident constituent), and +1st- and -1st-order diffracted light travel along the directions of the diffraction angle .theta. as shown (obliquely incident constituent). All these diffracted beams i.e., the zero-, +1st- and -1st-order diffracted light will interfere on the wafer and contribute to image formation.
The diffraction angle .theta. increases with finer mask patterns. If the sin.theta. is larger than NA, the +1st- and -1st-order diffracted light will not enter the projection lens, and then only zero-order diffracted light enters the projection lens to reach the surface of wafer, which then results in no interference. At this time, the minimum resolution is defined as: ##EQU2##
Meanwhile, since filter 3 is disposed in a position eccentric of the optical axis, the illumination light transmitted through a filter illuminates the mask having a specific obliquely incident angle, in the above obliquely incident illumination. Obliquely incident angle .alpha. is defined by a distance (x) between the optical axis and transmissive portion of the filter and by the focus length (f) of the condenser lens. EQU f sin (.alpha.)=x
The illumination light is diffracted by the mask pattern. The zero-order light is diffracted by angle .theta. with respect to the optical axis, and the angles between the path of +1st- and -1st-order diffracted light (.theta.1 and .theta.2, respectively) and the optical axis have the following relationships: EQU sin (.theta.1)+sin (.alpha.)=.lambda./Pr EQU sin (.theta.2)-sin (.alpha.)=.lambda./Pr
wherein, Pr is the line-to-space pitch of the mask.
Higher-order diffracted light travels other paths. Since the pitch of the pattern is fine and the sin(.theta.2) is larger than the NA of the projection lens in the mask side, -1st-order, and higher, diffracted light will not enter the projection lens. As a result, only zero-order and +1st-order diffracted light will interfere on the wafer surface and contribute to the image formation. The contrast of the image is approximately 90% when the mask patterns are lines and spaces with a duty of 1:1.
Here, the resolution (R) is defined as follows: ##EQU3##
Given a 5.times. projective magnification and a sin(.theta.1) of NA/2, the resolution limit on the wafer side is: ##EQU4##
The resolution limit is 1.5 times higher than that of a conventional projection exposure method. The DOF is also improved by the obliquely incident illumination method as described above.
In the above conventional oblique incident illumination method, given filter shapes as in FIGS. 2A or 2B, the area through which the light passes is much smaller than the area where light is blocked. In the case of FIG. 2A, the ratio of the light transmission is calculated by the expression (.sigma..sub.D.sup.2 -.sigma..sub.i.sup.2)/.sigma..sub.D.sup.2, and in general, .sigma..sub.i =2.sigma..sub.D /3 is known as the most preferable value wherein a transmission ratio is 5/9 and the exposure time is duplicated.
Referring to the case shown in FIG. 2B, the transmission ratio is calculated by the expression of 4.sigma..sub.i.sup.2 /.sigma..sub.D.sup.2 (where .sigma..sub.i =.sigma..sub.D/4), and the exposure ratio is 1/4, which quadruples the exposure time and the throughput is considerably reduced.
That is, since the conventional modified illumination system is a system comprised of the conventional illumination system and a filter attached thereto wherein the light transmitted to the illumination system is partially blocked, the exposure amount during the modified illumination is too small. Therefore, the exposure time becomes too long while error in the exposure system is generated, which results in the deterioration of uniformity.